(* This file is a part of IsarMathLib - a library of formalized mathematics for Isabelle/Isar. Copyright (C) 2005 - 2009 Slawomir Kolodynski This program is free software; Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. The name of the author may not be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. *) header{*\isaheader{Int\_ZF\_3.thy}*} theory Int_ZF_3 imports Int_ZF_2 begin text{*This theory is a continuation of @{text "Int_ZF_2"}. We consider here the properties of slopes (almost homomorphisms on integers) that allow to define the order relation and multiplicative inverse on real numbers. We also prove theorems that allow to show completeness of the order relation of real numbers we define in @{text "Real_ZF"}. *} section{*Positive slopes*} text{*This section provides background material for defining the order relation on real numbers.*} text{*Positive slopes are functions (of course.)*} lemma (in int1) Int_ZF_2_3_L1: assumes A1: "f∈\<S>⇩_{+}" shows "f:\<int>->\<int>" using assms AlmostHoms_def PositiveSet_def by simp text{*A small technical lemma to simplify the proof of the next theorem.*} lemma (in int1) Int_ZF_2_3_L1A: assumes A1: "f∈\<S>⇩_{+}" and A2: "∃n ∈ f``(\<int>⇩_{+}) ∩ \<int>⇩_{+}. a\<lsq>n" shows "∃M∈\<int>⇩_{+}. a \<lsq> f`(M)" proof - from A1 have "f:\<int>->\<int>" "\<int>⇩_{+}⊆ \<int>" using AlmostHoms_def PositiveSet_def by auto with A2 show ?thesis using func_imagedef by auto qed text{*The next lemma is Lemma 3 in the Arthan's paper.*} lemma (in int1) Arthan_Lem_3: assumes A1: "f∈\<S>⇩_{+}" and A2: "D ∈ \<int>⇩_{+}" shows "∃M∈\<int>⇩_{+}. ∀m∈\<int>⇩_{+}. (m\<ra>\<one>)·D \<lsq> f`(m·M)" proof - let ?E = "maxδ(f) \<ra> D" let ?A = "f``(\<int>⇩_{+}) ∩ \<int>⇩_{+}" from A1 A2 have I: "D\<lsq>?E" using Int_ZF_1_5_L3 Int_ZF_2_1_L8 Int_ZF_2_L1A Int_ZF_2_L15D by simp from A1 A2 have "?A ⊆ \<int>⇩_{+}" "?A ∉ Fin(\<int>)" "\<two>·?E ∈ \<int>" using int_two_three_are_int Int_ZF_2_1_L8 PositiveSet_def Int_ZF_1_1_L5 by auto with A1 have "∃M∈\<int>⇩_{+}. \<two>·?E \<lsq> f`(M)" using Int_ZF_1_5_L2A Int_ZF_2_3_L1A by simp then obtain M where II: "M∈\<int>⇩_{+}" and III: "\<two>·?E \<lsq> f`(M)" by auto { fix m assume "m∈\<int>⇩_{+}" then have A4: "\<one>\<lsq>m" using Int_ZF_1_5_L3 by simp moreover from II III have "(\<one>\<ra>\<one>) ·?E \<lsq> f`(\<one>·M)" using PositiveSet_def Int_ZF_1_1_L4 by simp moreover have "∀k. \<one>\<lsq>k ∧ (k\<ra>\<one>)·?E \<lsq> f`(k·M) --> (k\<ra>\<one>\<ra>\<one>)·?E \<lsq> f`((k\<ra>\<one>)·M)" proof - { fix k assume A5: "\<one>\<lsq>k" and A6: "(k\<ra>\<one>)·?E \<lsq> f`(k·M)" with A1 A2 II have T: "k∈\<int>" "M∈\<int>" "k\<ra>\<one> ∈ \<int>" "?E∈\<int>" "(k\<ra>\<one>)·?E ∈ \<int>" "\<two>·?E ∈ \<int>" using Int_ZF_2_L1A PositiveSet_def int_zero_one_are_int Int_ZF_1_1_L5 Int_ZF_2_1_L8 by auto from A1 A2 A5 II have "δ(f,k·M,M) ∈ \<int>" "abs(δ(f,k·M,M)) \<lsq> maxδ(f)" "\<zero>\<lsq>D" using Int_ZF_2_L1A PositiveSet_def Int_ZF_1_1_L5 Int_ZF_2_1_L7 Int_ZF_2_L16C by auto with III A6 have "(k\<ra>\<one>)·?E \<ra> (\<two>·?E \<rs> ?E) \<lsq> f`(k·M) \<ra> (f`(M) \<ra> δ(f,k·M,M))" using Int_ZF_1_3_L19A int_ineq_add_sides by simp with A1 T have "(k\<ra>\<one>\<ra>\<one>)·?E \<lsq> f`((k\<ra>\<one>)·M)" using Int_ZF_1_1_L1 int_zero_one_are_int Int_ZF_1_1_L4 Int_ZF_1_2_L11 Int_ZF_2_1_L13 by simp } then show ?thesis by simp qed ultimately have "(m\<ra>\<one>)·?E \<lsq> f`(m·M)" by (rule Induction_on_int) with A4 I have "(m\<ra>\<one>)·D \<lsq> f`(m·M)" using Int_ZF_1_3_L13A by simp } then have "∀m∈\<int>⇩_{+}.(m\<ra>\<one>)·D \<lsq> f`(m·M)" by simp with II show ?thesis by auto qed text{*A special case of @{text " Arthan_Lem_3"} when $D=1$.*} corollary (in int1) Arthan_L_3_spec: assumes A1: "f ∈ \<S>⇩_{+}" shows "∃M∈\<int>⇩_{+}.∀n∈\<int>⇩_{+}. n\<ra>\<one> \<lsq> f`(n·M)" proof - have "∀n∈\<int>⇩_{+}. n\<ra>\<one> ∈ \<int>" using PositiveSet_def int_zero_one_are_int Int_ZF_1_1_L5 by simp then have "∀n∈\<int>⇩_{+}. (n\<ra>\<one>)·\<one> = n\<ra>\<one>" using Int_ZF_1_1_L4 by simp moreover from A1 have "∃M∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. (n\<ra>\<one>)·\<one> \<lsq> f`(n·M)" using int_one_two_are_pos Arthan_Lem_3 by simp ultimately show ?thesis by simp qed text{*We know from @{text "Group_ZF_3.thy"} that finite range functions are almost homomorphisms. Besides reminding that fact for slopes the next lemma shows that finite range functions do not belong to @{text "\<S>⇩_{+}"}. This is important, because the projection of the set of finite range functions defines zero in the real number construction in @{text "Real_ZF_x.thy"} series, while the projection of @{text "\<S>⇩_{+}"} becomes the set of (strictly) positive reals. We don't want zero to be positive, do we? The next lemma is a part of Lemma 5 in the Arthan's paper \cite{Arthan2004}.*} lemma (in int1) Int_ZF_2_3_L1B: assumes A1: "f ∈ FinRangeFunctions(\<int>,\<int>)" shows "f∈\<S>" "f ∉ \<S>⇩_{+}" proof - from A1 show "f∈\<S>" using Int_ZF_2_1_L1 group1.Group_ZF_3_3_L1 by auto have "\<int>⇩_{+}⊆ \<int>" using PositiveSet_def by auto with A1 have "f``(\<int>⇩_{+}) ∈ Fin(\<int>)" using Finite1_L21 by simp then have "f``(\<int>⇩_{+}) ∩ \<int>⇩_{+}∈ Fin(\<int>)" using Fin_subset_lemma by blast thus "f ∉ \<S>⇩_{+}" by auto qed text{*We want to show that if $f$ is a slope and neither $f$ nor $-f$ are in @{text "\<S>⇩_{+}"}, then $f$ is bounded. The next lemma is the first step towards that goal and shows that if slope is not in @{text "\<S>⇩_{+}"} then $f($@{text "\<int>⇩_{+}"}$)$ is bounded above.*} lemma (in int1) Int_ZF_2_3_L2: assumes A1: "f∈\<S>" and A2: "f ∉ \<S>⇩_{+}" shows "IsBoundedAbove(f``(\<int>⇩_{+}), IntegerOrder)" proof - from A1 have "f:\<int>->\<int>" using AlmostHoms_def by simp then have "f``(\<int>⇩_{+}) ⊆ \<int>" using func1_1_L6 by simp moreover from A1 A2 have "f``(\<int>⇩_{+}) ∩ \<int>⇩_{+}∈ Fin(\<int>)" by auto ultimately show ?thesis using Int_ZF_2_T1 group3.OrderedGroup_ZF_2_L4 by simp qed text{*If $f$ is a slope and $-f\notin$ @{text "\<S>⇩_{+}"}, then $f($@{text "\<int>⇩_{+}"}$)$ is bounded below.*} lemma (in int1) Int_ZF_2_3_L3: assumes A1: "f∈\<S>" and A2: "\<fm>f ∉ \<S>⇩_{+}" shows "IsBoundedBelow(f``(\<int>⇩_{+}), IntegerOrder)" proof - from A1 have T: "f:\<int>->\<int>" using AlmostHoms_def by simp then have "(\<sm>(f``(\<int>⇩_{+}))) = (\<fm>f)``(\<int>⇩_{+})" using Int_ZF_1_T2 group0_2_T2 PositiveSet_def func1_1_L15C by auto with A1 A2 T show "IsBoundedBelow(f``(\<int>⇩_{+}), IntegerOrder)" using Int_ZF_2_1_L12 Int_ZF_2_3_L2 PositiveSet_def func1_1_L6 Int_ZF_2_T1 group3.OrderedGroup_ZF_2_L5 by simp qed text{* A slope that is bounded on @{text "\<int>⇩_{+}"} is bounded everywhere.*} lemma (in int1) Int_ZF_2_3_L4: assumes A1: "f∈\<S>" and A2: "m∈\<int>" and A3: "∀n∈\<int>⇩_{+}. abs(f`(n)) \<lsq> L" shows "abs(f`(m)) \<lsq> \<two>·maxδ(f) \<ra> L" proof - from A1 A3 have "\<zero> \<lsq> abs(f`(\<one>))" "abs(f`(\<one>)) \<lsq> L" using int_zero_one_are_int Int_ZF_2_1_L2B int_abs_nonneg int_one_two_are_pos by auto then have II: "\<zero>\<lsq>L" by (rule Int_order_transitive) note A2 moreover have "abs(f`(\<zero>)) \<lsq> \<two>·maxδ(f) \<ra> L" proof - from A1 have "abs(f`(\<zero>)) \<lsq> maxδ(f)" "\<zero> \<lsq> maxδ(f)" and T: "maxδ(f) ∈ \<int>" using Int_ZF_2_1_L8 by auto with II have "abs(f`(\<zero>)) \<lsq> maxδ(f) \<ra> maxδ(f) \<ra> L" using Int_ZF_2_L15F by simp with T show ?thesis using Int_ZF_1_1_L4 by simp qed moreover from A1 A3 II have "∀n∈\<int>⇩_{+}. abs(f`(n)) \<lsq> \<two>·maxδ(f) \<ra> L" using Int_ZF_2_1_L8 Int_ZF_1_3_L5A Int_ZF_2_L15F by simp moreover have "∀n∈\<int>⇩_{+}. abs(f`(\<rm>n)) \<lsq> \<two>·maxδ(f) \<ra> L" proof fix n assume "n∈\<int>⇩_{+}" with A1 A3 have "\<two>·maxδ(f) ∈ \<int>" "abs(f`(\<rm>n)) \<lsq> \<two>·maxδ(f) \<ra> abs(f`(n))" "abs(f`(n)) \<lsq> L" using int_two_three_are_int Int_ZF_2_1_L8 Int_ZF_1_1_L5 PositiveSet_def Int_ZF_2_1_L14 by auto then show "abs(f`(\<rm>n)) \<lsq> \<two>·maxδ(f) \<ra> L" using Int_ZF_2_L15A by blast qed ultimately show ?thesis by (rule Int_ZF_2_L19B) qed text{*A slope whose image of the set of positive integers is bounded is a finite range function.*} lemma (in int1) Int_ZF_2_3_L4A: assumes A1: "f∈\<S>" and A2: "IsBounded(f``(\<int>⇩_{+}), IntegerOrder)" shows "f ∈ FinRangeFunctions(\<int>,\<int>)" proof - have T1: "\<int>⇩_{+}⊆ \<int>" using PositiveSet_def by auto from A1 have T2: "f:\<int>->\<int>" using AlmostHoms_def by simp from A2 obtain L where "∀a∈f``(\<int>⇩_{+}). abs(a) \<lsq> L" using Int_ZF_1_3_L20A by auto with T2 T1 have "∀n∈\<int>⇩_{+}. abs(f`(n)) \<lsq> L" by (rule func1_1_L15B) with A1 have "∀m∈\<int>. abs(f`(m)) \<lsq> \<two>·maxδ(f) \<ra> L" using Int_ZF_2_3_L4 by simp with T2 have "f``(\<int>) ∈ Fin(\<int>)" by (rule Int_ZF_1_3_L20C) with T2 show "f ∈ FinRangeFunctions(\<int>,\<int>)" using FinRangeFunctions_def by simp qed text{*A slope whose image of the set of positive integers is bounded below is a finite range function or a positive slope.*} lemma (in int1) Int_ZF_2_3_L4B: assumes "f∈\<S>" and "IsBoundedBelow(f``(\<int>⇩_{+}), IntegerOrder)" shows "f ∈ FinRangeFunctions(\<int>,\<int>) ∨ f∈\<S>⇩_{+}" using assms Int_ZF_2_3_L2 IsBounded_def Int_ZF_2_3_L4A by auto text{*If one slope is not greater then another on positive integers, then they are almost equal or the difference is a positive slope.*} lemma (in int1) Int_ZF_2_3_L4C: assumes A1: "f∈\<S>" "g∈\<S>" and A2: "∀n∈\<int>⇩_{+}. f`(n) \<lsq> g`(n)" shows "f∼g ∨ g \<fp> (\<fm>f) ∈ \<S>⇩_{+}" proof - let ?h = "g \<fp> (\<fm>f)" from A1 have "(\<fm>f) ∈ \<S>" using Int_ZF_2_1_L12 by simp with A1 have I: "?h ∈ \<S>" using Int_ZF_2_1_L12C by simp moreover have "IsBoundedBelow(?h``(\<int>⇩_{+}), IntegerOrder)" proof - from I have "?h:\<int>->\<int>" and "\<int>⇩_{+}⊆\<int>" using AlmostHoms_def PositiveSet_def by auto moreover from A1 A2 have "∀n∈\<int>⇩_{+}. ⟨\<zero>, ?h`(n)⟩ ∈ IntegerOrder" using Int_ZF_2_1_L2B PositiveSet_def Int_ZF_1_3_L10A Int_ZF_2_1_L12 Int_ZF_2_1_L12B Int_ZF_2_1_L12A by simp ultimately show "IsBoundedBelow(?h``(\<int>⇩_{+}), IntegerOrder)" by (rule func_ZF_8_L1) qed ultimately have "?h ∈ FinRangeFunctions(\<int>,\<int>) ∨ ?h∈\<S>⇩_{+}" using Int_ZF_2_3_L4B by simp with A1 show "f∼g ∨ g \<fp> (\<fm>f) ∈ \<S>⇩_{+}" using Int_ZF_2_1_L9C by auto qed text{*Positive slopes are arbitrarily large for large enough arguments.*} lemma (in int1) Int_ZF_2_3_L5: assumes A1: "f∈\<S>⇩_{+}" and A2: "K∈\<int>" shows "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> K \<lsq> f`(m)" proof - from A1 obtain M where I: "M∈\<int>⇩_{+}" and II: "∀n∈\<int>⇩_{+}. n\<ra>\<one> \<lsq> f`(n·M)" using Arthan_L_3_spec by auto let ?j = "GreaterOf(IntegerOrder,M,K \<rs> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f)) \<rs> \<one>)" from A1 I have T1: "minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f) ∈ \<int>" "M∈\<int>" using Int_ZF_2_1_L15 Int_ZF_2_1_L8 Int_ZF_1_1_L5 PositiveSet_def by auto with A2 I have T2: "K \<rs> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f)) ∈ \<int>" "K \<rs> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f)) \<rs> \<one> ∈ \<int>" using Int_ZF_1_1_L5 int_zero_one_are_int by auto with T1 have III: "M \<lsq> ?j" and "K \<rs> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f)) \<rs> \<one> \<lsq> ?j" using Int_ZF_1_3_L18 by auto with A2 T1 T2 have IV: "K \<lsq> ?j\<ra>\<one> \<ra> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f))" using int_zero_one_are_int Int_ZF_2_L9C by simp let ?N = "GreaterOf(IntegerOrder,\<one>,?j·M)" from T1 III have T3: "?j ∈ \<int>" "?j·M ∈ \<int>" using Int_ZF_2_L1A Int_ZF_1_1_L5 by auto then have V: "?N ∈ \<int>⇩_{+}" and VI: "?j·M \<lsq> ?N" using int_zero_one_are_int Int_ZF_1_5_L3 Int_ZF_1_3_L18 by auto { fix m let ?n = "m zdiv M" let ?k = "m zmod M" assume "?N\<lsq>m" with VI have "?j·M \<lsq> m" by (rule Int_order_transitive) with I III have VII: "m = ?n·M\<ra>?k" "?j \<lsq> ?n" and VIII: "?n ∈ \<int>⇩_{+}" "?k ∈ \<zero>..(M\<rs>\<one>)" using IntDiv_ZF_1_L5 by auto with II have "?j \<ra> \<one> \<lsq> ?n \<ra> \<one>" "?n\<ra>\<one> \<lsq> f`(?n·M)" using int_zero_one_are_int int_ord_transl_inv by auto then have "?j \<ra> \<one> \<lsq> f`(?n·M)" by (rule Int_order_transitive) with T1 have "?j\<ra>\<one> \<ra> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f)) \<lsq> f`(?n·M) \<ra> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f))" using int_ord_transl_inv by simp with IV have "K \<lsq> f`(?n·M) \<ra> (minf(f,\<zero>..(M\<rs>\<one>)) \<rs> maxδ(f))" by (rule Int_order_transitive) moreover from A1 I VIII have "f`(?n·M) \<ra> (minf(f,\<zero>..(M\<rs>\<one>))\<rs> maxδ(f)) \<lsq> f`(?n·M\<ra>?k)" using PositiveSet_def Int_ZF_2_1_L16 by simp ultimately have "K \<lsq> f`(?n·M\<ra>?k)" by (rule Int_order_transitive) with VII have "K \<lsq> f`(m)" by simp } then have "∀m. ?N\<lsq>m --> K \<lsq> f`(m)" by simp with V show ?thesis by auto qed text{*Positive slopes are arbitrarily small for small enough arguments. Kind of dual to @{text "Int_ZF_2_3_L5"}.*} lemma (in int1) Int_ZF_2_3_L5A: assumes A1: "f∈\<S>⇩_{+}" and A2: "K∈\<int>" shows "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> f`(\<rm>m) \<lsq> K" proof - from A1 have T1: "abs(f`(\<zero>)) \<ra> maxδ(f) ∈ \<int>" using Int_ZF_2_1_L8 by auto with A2 have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K ∈ \<int>" using Int_ZF_1_1_L5 by simp with A1 have "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> f`(m)" using Int_ZF_2_3_L5 by simp then obtain N where I: "N∈\<int>⇩_{+}" and II: "∀m. N\<lsq>m --> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> f`(m)" by auto { fix m assume A3: "N\<lsq>m" with A1 have "f`(\<rm>m) \<lsq> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> f`(m)" using Int_ZF_2_L1A Int_ZF_2_1_L14 by simp moreover from II T1 A3 have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> f`(m) \<lsq> (abs(f`(\<zero>)) \<ra> maxδ(f)) \<rs>(abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K)" using Int_ZF_2_L10 int_ord_transl_inv by simp with A2 T1 have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> f`(m) \<lsq> K" using Int_ZF_1_2_L3 by simp ultimately have "f`(\<rm>m) \<lsq> K" by (rule Int_order_transitive) } then have "∀m. N\<lsq>m --> f`(\<rm>m) \<lsq> K" by simp with I show ?thesis by auto qed (*lemma (in int1) Int_ZF_2_3_L5A: assumes A1: "f∈\<S>⇩_{+}" and A2: "K∈\<int>" shows "∃N∈\<int>⇩_{+}. ∀m. m\<lsq>(\<rm>N) --> f`(m) \<lsq> K" proof - from A1 have T1: "abs(f`(\<zero>)) \<ra> maxδ(f) ∈ \<int>" using Int_ZF_2_1_L8 by auto; with A2 have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K ∈ \<int>" using Int_ZF_1_1_L5 by simp; with A1 have "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> f`(m)" using Int_ZF_2_3_L5 by simp; then obtain N where I: "N∈\<int>⇩_{+}" and II: "∀m. N\<lsq>m --> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> f`(m)" by auto; { fix m assume A3: "m\<lsq>(\<rm>N)" with A1 have T2: "f`(m) ∈ \<int>" using Int_ZF_2_L1A Int_ZF_2_1_L2B by simp; from A1 I II A3 have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> f`(\<rm>m)" and "f`(\<rm>m) \<lsq> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> f`(m)" using PositiveSet_def Int_ZF_2_L10AA Int_ZF_2_L1A Int_ZF_2_1_L14 by auto; then have "abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> K \<lsq> abs(f`(\<zero>)) \<ra> maxδ(f) \<rs> f`(m)" by (rule Int_order_transitive) with T1 A2 T2 have "f`(m) \<lsq> K" using Int_ZF_2_L10AB by simp; } then have "∀m. m\<lsq>(\<rm>N) --> f`(m) \<lsq> K" by simp; with I show ?thesis by auto; qed;*) text{*A special case of @{text "Int_ZF_2_3_L5"} where $K=1$.*} corollary (in int1) Int_ZF_2_3_L6: assumes "f∈\<S>⇩_{+}" shows "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> f`(m) ∈ \<int>⇩_{+}" using assms int_zero_one_are_int Int_ZF_2_3_L5 Int_ZF_1_5_L3 by simp text{*A special case of @{text "Int_ZF_2_3_L5"} where $m=N$.*} corollary (in int1) Int_ZF_2_3_L6A: assumes "f∈\<S>⇩_{+}" and "K∈\<int>" shows "∃N∈\<int>⇩_{+}. K \<lsq> f`(N)" proof - from assms have "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> K \<lsq> f`(m)" using Int_ZF_2_3_L5 by simp then obtain N where I: "N ∈ \<int>⇩_{+}" and II: "∀m. N\<lsq>m --> K \<lsq> f`(m)" by auto then show ?thesis using PositiveSet_def int_ord_is_refl refl_def by auto qed text{*If values of a slope are not bounded above, then the slope is positive.*} lemma (in int1) Int_ZF_2_3_L7: assumes A1: "f∈\<S>" and A2: "∀K∈\<int>. ∃n∈\<int>⇩_{+}. K \<lsq> f`(n)" shows "f ∈ \<S>⇩_{+}" proof - { fix K assume "K∈\<int>" with A2 obtain n where "n∈\<int>⇩_{+}" "K \<lsq> f`(n)" by auto moreover from A1 have "\<int>⇩_{+}⊆ \<int>" "f:\<int>->\<int>" using PositiveSet_def AlmostHoms_def by auto ultimately have "∃m ∈ f``(\<int>⇩_{+}). K \<lsq> m" using func1_1_L15D by auto } then have "∀K∈\<int>. ∃m ∈ f``(\<int>⇩_{+}). K \<lsq> m" by simp with A1 show "f ∈ \<S>⇩_{+}" using Int_ZF_4_L9 Int_ZF_2_3_L2 by auto qed text{*For unbounded slope $f$ either $f\in$@{text "\<S>⇩_{+}"} of $-f\in$@{text "\<S>⇩_{+}"}.*} theorem (in int1) Int_ZF_2_3_L8: assumes A1: "f∈\<S>" and A2: "f ∉ FinRangeFunctions(\<int>,\<int>)" shows "(f ∈ \<S>⇩_{+}) Xor ((\<fm>f) ∈ \<S>⇩_{+})" proof - have T1: "\<int>⇩_{+}⊆ \<int>" using PositiveSet_def by auto from A1 have T2: "f:\<int>->\<int>" using AlmostHoms_def by simp then have I: "f``(\<int>⇩_{+}) ⊆ \<int>" using func1_1_L6 by auto from A1 A2 have "f ∈ \<S>⇩_{+}∨ (\<fm>f) ∈ \<S>⇩_{+}" using Int_ZF_2_3_L2 Int_ZF_2_3_L3 IsBounded_def Int_ZF_2_3_L4A by blast moreover have "¬(f ∈ \<S>⇩_{+}∧ (\<fm>f) ∈ \<S>⇩_{+})" proof - { assume A3: "f ∈ \<S>⇩_{+}" and A4: "(\<fm>f) ∈ \<S>⇩_{+}" from A3 obtain N1 where I: "N1∈\<int>⇩_{+}" and II: "∀m. N1\<lsq>m --> f`(m) ∈ \<int>⇩_{+}" using Int_ZF_2_3_L6 by auto from A4 obtain N2 where III: "N2∈\<int>⇩_{+}" and IV: "∀m. N2\<lsq>m --> (\<fm>f)`(m) ∈ \<int>⇩_{+}" using Int_ZF_2_3_L6 by auto let ?N = "GreaterOf(IntegerOrder,N1,N2)" from I III have "N1 \<lsq> ?N" "N2 \<lsq> ?N" using PositiveSet_def Int_ZF_1_3_L18 by auto with A1 II IV have "f`(?N) ∈ \<int>⇩_{+}" "(\<fm>f)`(?N) ∈ \<int>⇩_{+}" "(\<fm>f)`(?N) = \<rm>(f`(?N))" using Int_ZF_2_L1A PositiveSet_def Int_ZF_2_1_L12A by auto then have False using Int_ZF_1_5_L8 by simp } thus ?thesis by auto qed ultimately show "(f ∈ \<S>⇩_{+}) Xor ((\<fm>f) ∈ \<S>⇩_{+})" using Xor_def by simp qed text{*The sum of positive slopes is a positive slope.*} theorem (in int1) sum_of_pos_sls_is_pos_sl: assumes A1: "f ∈ \<S>⇩_{+}" "g ∈ \<S>⇩_{+}" shows "f\<fp>g ∈ \<S>⇩_{+}" proof - { fix K assume "K∈\<int>" with A1 have "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> K \<lsq> f`(m)" using Int_ZF_2_3_L5 by simp then obtain N where I: "N∈\<int>⇩_{+}" and II: "∀m. N\<lsq>m --> K \<lsq> f`(m)" by auto from A1 have "∃M∈\<int>⇩_{+}. ∀m. M\<lsq>m --> \<zero> \<lsq> g`(m)" using int_zero_one_are_int Int_ZF_2_3_L5 by simp then obtain M where III: "M∈\<int>⇩_{+}" and IV: "∀m. M\<lsq>m --> \<zero> \<lsq> g`(m)" by auto let ?L = "GreaterOf(IntegerOrder,N,M)" from I III have V: "?L ∈ \<int>⇩_{+}" "\<int>⇩_{+}⊆ \<int>" using GreaterOf_def PositiveSet_def by auto moreover from A1 V have "(f\<fp>g)`(?L) = f`(?L) \<ra> g`(?L)" using Int_ZF_2_1_L12B by auto moreover from I II III IV have "K \<lsq> f`(?L) \<ra> g`(?L)" using PositiveSet_def Int_ZF_1_3_L18 Int_ZF_2_L15F by simp ultimately have "?L ∈ \<int>⇩_{+}" "K \<lsq> (f\<fp>g)`(?L)" by auto then have "∃n ∈\<int>⇩_{+}. K \<lsq> (f\<fp>g)`(n)" by auto } with A1 show "f\<fp>g ∈ \<S>⇩_{+}" using Int_ZF_2_1_L12C Int_ZF_2_3_L7 by simp qed text{*The composition of positive slopes is a positive slope.*} theorem (in int1) comp_of_pos_sls_is_pos_sl: assumes A1: "f ∈ \<S>⇩_{+}" "g ∈ \<S>⇩_{+}" shows "fog ∈ \<S>⇩_{+}" proof - { fix K assume "K∈\<int>" with A1 have "∃N∈\<int>⇩_{+}. ∀m. N\<lsq>m --> K \<lsq> f`(m)" using Int_ZF_2_3_L5 by simp then obtain N where "N∈\<int>⇩_{+}" and I: "∀m. N\<lsq>m --> K \<lsq> f`(m)" by auto with A1 have "∃M∈\<int>⇩_{+}. N \<lsq> g`(M)" using PositiveSet_def Int_ZF_2_3_L6A by simp then obtain M where "M∈\<int>⇩_{+}" "N \<lsq> g`(M)" by auto with A1 I have "∃M∈\<int>⇩_{+}. K \<lsq> (fog)`(M)" using PositiveSet_def Int_ZF_2_1_L10 by auto } with A1 show "fog ∈ \<S>⇩_{+}" using Int_ZF_2_1_L11 Int_ZF_2_3_L7 by simp qed text{*A slope equivalent to a positive one is positive.*} lemma (in int1) Int_ZF_2_3_L9: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "⟨f,g⟩ ∈ AlEqRel" shows "g ∈ \<S>⇩_{+}" proof - from A2 have T: "g∈\<S>" and "∃L∈\<int>. ∀m∈\<int>. abs(f`(m)\<rs>g`(m)) \<lsq> L" using Int_ZF_2_1_L9A by auto then obtain L where I: "L∈\<int>" and II: "∀m∈\<int>. abs(f`(m)\<rs>g`(m)) \<lsq> L" by auto { fix K assume A3: "K∈\<int>" with I have "K\<ra>L ∈ \<int>" using Int_ZF_1_1_L5 by simp with A1 obtain M where III: "M∈\<int>⇩_{+}" and IV: "K\<ra>L \<lsq> f`(M)" using Int_ZF_2_3_L6A by auto with A1 A3 I have "K \<lsq> f`(M)\<rs>L" using PositiveSet_def Int_ZF_2_1_L2B Int_ZF_2_L9B by simp moreover from A1 T II III have "f`(M)\<rs>L \<lsq> g`(M)" using PositiveSet_def Int_ZF_2_1_L2B Int_triangle_ineq2 by simp ultimately have "K \<lsq> g`(M)" by (rule Int_order_transitive) with III have "∃n∈\<int>⇩_{+}. K \<lsq> g`(n)" by auto } with T show "g ∈ \<S>⇩_{+}" using Int_ZF_2_3_L7 by simp qed text{* The set of positive slopes is saturated with respect to the relation of equivalence of slopes.*} lemma (in int1) pos_slopes_saturated: shows "IsSaturated(AlEqRel,\<S>⇩_{+})" proof - have "equiv(\<S>,AlEqRel)" "AlEqRel ⊆ \<S> × \<S>" using Int_ZF_2_1_L9B by auto moreover have "\<S>⇩_{+}⊆ \<S>" by auto moreover have "∀f∈\<S>⇩_{+}. ∀g∈\<S>. ⟨f,g⟩ ∈ AlEqRel --> g ∈ \<S>⇩_{+}" using Int_ZF_2_3_L9 by blast ultimately show "IsSaturated(AlEqRel,\<S>⇩_{+})" by (rule EquivClass_3_L3) qed text{*A technical lemma involving a projection of the set of positive slopes and a logical epression with exclusive or.*} lemma (in int1) Int_ZF_2_3_L10: assumes A1: "f∈\<S>" "g∈\<S>" and A2: "R = {AlEqRel``{s}. s∈\<S>⇩_{+}}" and A3: "(f∈\<S>⇩_{+}) Xor (g∈\<S>⇩_{+})" shows "(AlEqRel``{f} ∈ R) Xor (AlEqRel``{g} ∈ R)" proof - from A1 A2 A3 have "equiv(\<S>,AlEqRel)" "IsSaturated(AlEqRel,\<S>⇩_{+})" "\<S>⇩_{+}⊆ \<S>" "f∈\<S>" "g∈\<S>" "R = {AlEqRel``{s}. s∈\<S>⇩_{+}}" "(f∈\<S>⇩_{+}) Xor (g∈\<S>⇩_{+})" using pos_slopes_saturated Int_ZF_2_1_L9B by auto then show ?thesis by (rule EquivClass_3_L7) qed text{*Identity function is a positive slope.*} lemma (in int1) Int_ZF_2_3_L11: shows "id(\<int>) ∈ \<S>⇩_{+}" proof - let ?f = "id(\<int>)" { fix K assume "K∈\<int>" then obtain n where T: "n∈\<int>⇩_{+}" and "K\<lsq>n" using Int_ZF_1_5_L9 by auto moreover from T have "?f`(n) = n" using PositiveSet_def by simp ultimately have "n∈\<int>⇩_{+}" and "K\<lsq>?f`(n)" by auto then have "∃n∈\<int>⇩_{+}. K\<lsq>?f`(n)" by auto } then show "?f ∈ \<S>⇩_{+}" using Int_ZF_2_1_L17 Int_ZF_2_3_L7 by simp qed text{*The identity function is not almost equal to any bounded function.*} lemma (in int1) Int_ZF_2_3_L12: assumes A1: "f ∈ FinRangeFunctions(\<int>,\<int>)" shows "¬(id(\<int>) ∼ f)" proof - { from A1 have "id(\<int>) ∈ \<S>⇩_{+}" using Int_ZF_2_3_L11 by simp moreover assume "⟨id(\<int>),f⟩ ∈ AlEqRel" ultimately have "f ∈ \<S>⇩_{+}" by (rule Int_ZF_2_3_L9) with A1 have False using Int_ZF_2_3_L1B by simp } then show "¬(id(\<int>) ∼ f)" by auto qed section{*Inverting slopes*} text{*Not every slope is a 1:1 function. However, we can still invert slopes in the sense that if $f$ is a slope, then we can find a slope $g$ such that $f\circ g$ is almost equal to the identity function. The goal of this this section is to establish this fact for positive slopes. *} text{*If $f$ is a positive slope, then for every positive integer $p$ the set $\{n\in Z_+: p\leq f(n)\}$ is a nonempty subset of positive integers. Recall that $f^{-1}(p)$ is the notation for the smallest element of this set.*} lemma (in int1) Int_ZF_2_4_L1: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "p∈\<int>⇩_{+}" and A3: "A = {n∈\<int>⇩_{+}. p \<lsq> f`(n)}" shows "A ⊆ \<int>⇩_{+}" "A ≠ 0" "f¯(p) ∈ A" "∀m∈A. f¯(p) \<lsq> m" proof - from A3 show I: "A ⊆ \<int>⇩_{+}" by auto from A1 A2 have "∃n∈\<int>⇩_{+}. p \<lsq> f`(n)" using PositiveSet_def Int_ZF_2_3_L6A by simp with A3 show II: "A ≠ 0" by auto from A3 I II show "f¯(p) ∈ A" "∀m∈A. f¯(p) \<lsq> m" using Int_ZF_1_5_L1C by auto qed text{*If $f$ is a positive slope and $p$ is a positive integer $p$, then $f^{-1}(p)$ (defined as the minimum of the set $\{n\in Z_+: p\leq f(n)\}$ ) is a (well defined) positive integer.*} lemma (in int1) Int_ZF_2_4_L2: assumes "f ∈ \<S>⇩_{+}" and "p∈\<int>⇩_{+}" shows "f¯(p) ∈ \<int>⇩_{+}" "p \<lsq> f`(f¯(p))" using assms Int_ZF_2_4_L1 by auto text{*If $f$ is a positive slope and $p$ is a positive integer such that $n\leq f(p)$, then $f^{-1}(n) \leq p$.*} lemma (in int1) Int_ZF_2_4_L3: assumes "f ∈ \<S>⇩_{+}" and "m∈\<int>⇩_{+}" "p∈\<int>⇩_{+}" and "m \<lsq> f`(p)" shows "f¯(m) \<lsq> p" using assms Int_ZF_2_4_L1 by simp text{*An upper bound $f(f^{-1}(m) -1)$ for positive slopes.*} lemma (in int1) Int_ZF_2_4_L4: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "m∈\<int>⇩_{+}" and A3: "f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" shows "f`(f¯(m)\<rs>\<one>) \<lsq> m" "f`(f¯(m)\<rs>\<one>) ≠ m" proof - from A1 A2 have T: "f¯(m) ∈ \<int>" using Int_ZF_2_4_L2 PositiveSet_def by simp from A1 A3 have "f:\<int>->\<int>" and "f¯(m)\<rs>\<one> ∈ \<int>" using Int_ZF_2_3_L1 PositiveSet_def by auto with A1 A2 have T1: "f`(f¯(m)\<rs>\<one>) ∈ \<int>" "m∈\<int>" using apply_funtype PositiveSet_def by auto { assume "m \<lsq> f`(f¯(m)\<rs>\<one>)" with A1 A2 A3 have "f¯(m) \<lsq> f¯(m)\<rs>\<one>" by (rule Int_ZF_2_4_L3) with T have False using Int_ZF_1_2_L3AA by simp } then have I: "¬(m \<lsq> f`(f¯(m)\<rs>\<one>))" by auto with T1 show "f`(f¯(m)\<rs>\<one>) \<lsq> m" by (rule Int_ZF_2_L19) from T1 I show "f`(f¯(m)\<rs>\<one>) ≠ m" by (rule Int_ZF_2_L19) qed text{*The (candidate for) the inverse of a positive slope is nondecreasing.*} lemma (in int1) Int_ZF_2_4_L5: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "m∈\<int>⇩_{+}" and A3: "m\<lsq>n" shows "f¯(m) \<lsq> f¯(n)" proof - from A2 A3 have T: "n ∈ \<int>⇩_{+}" using Int_ZF_1_5_L7 by blast with A1 have "n \<lsq> f`(f¯(n))" using Int_ZF_2_4_L2 by simp with A3 have "m \<lsq> f`(f¯(n))" by (rule Int_order_transitive) with A1 A2 T show "f¯(m) \<lsq> f¯(n)" using Int_ZF_2_4_L2 Int_ZF_2_4_L3 by simp qed text{*If $f^{-1}(m)$ is positive and $n$ is a positive integer, then, then $f^{-1}(m+n)-1$ is positive.*} lemma (in int1) Int_ZF_2_4_L6: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "m∈\<int>⇩_{+}" "n∈\<int>⇩_{+}" and A3: "f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" shows "f¯(m\<ra>n)\<rs>\<one> ∈ \<int>⇩_{+}" proof - from A1 A2 have "f¯(m)\<rs>\<one> \<lsq> f¯(m\<ra>n) \<rs> \<one>" using PositiveSet_def Int_ZF_1_5_L7A Int_ZF_2_4_L2 Int_ZF_2_4_L5 int_zero_one_are_int Int_ZF_1_1_L4 int_ord_transl_inv by simp with A3 show "f¯(m\<ra>n)\<rs>\<one> ∈ \<int>⇩_{+}" using Int_ZF_1_5_L7 by blast qed text{*If $f$ is a slope, then $f(f^{-1}(m+n)-f^{-1}(m) - f^{-1}(n))$ is uniformly bounded above and below. Will it be the messiest IsarMathLib proof ever? Only time will tell.*} lemma (in int1) Int_ZF_2_4_L7: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "∀m∈\<int>⇩_{+}. f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" shows "∃U∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n)) \<lsq> U" "∃N∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. N \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" proof - from A1 have "∃L∈\<int>. ∀r∈\<int>. f`(r) \<lsq> f`(r\<rs>\<one>) \<ra> L" using Int_ZF_2_1_L28 by simp then obtain L where I: "L∈\<int>" and II: "∀r∈\<int>. f`(r) \<lsq> f`(r\<rs>\<one>) \<ra> L" by auto from A1 have "∃M∈\<int>. ∀r∈\<int>.∀p∈\<int>.∀q∈\<int>. f`(r\<rs>p\<rs>q) \<lsq> f`(r)\<rs>f`(p)\<rs>f`(q)\<ra>M" "∃K∈\<int>. ∀r∈\<int>.∀p∈\<int>.∀q∈\<int>. f`(r)\<rs>f`(p)\<rs>f`(q)\<ra>K \<lsq> f`(r\<rs>p\<rs>q)" using Int_ZF_2_1_L30 by auto then obtain M K where III: "M∈\<int>" and IV: "∀r∈\<int>.∀p∈\<int>.∀q∈\<int>. f`(r\<rs>p\<rs>q) \<lsq> f`(r)\<rs>f`(p)\<rs>f`(q)\<ra>M" and V: "K∈\<int>" and VI: "∀r∈\<int>.∀p∈\<int>.∀q∈\<int>. f`(r)\<rs>f`(p)\<rs>f`(q)\<ra>K \<lsq> f`(r\<rs>p\<rs>q)" by auto from I III V have "L\<ra>M ∈ \<int>" "(\<rm>L) \<rs> L \<ra> K ∈ \<int>" using Int_ZF_1_1_L4 Int_ZF_1_1_L5 by auto moreover { fix m n assume A3: "m∈\<int>⇩_{+}" "n∈\<int>⇩_{+}" have "f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n)) \<lsq> L\<ra>M ∧ (\<rm>L)\<rs>L\<ra>K \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" proof - let ?r = "f¯(m\<ra>n)" let ?p = "f¯(m)" let ?q = "f¯(n)" from A1 A3 have T1: "?p ∈ \<int>⇩_{+}" "?q ∈ \<int>⇩_{+}" "?r ∈ \<int>⇩_{+}" using Int_ZF_2_4_L2 pos_int_closed_add_unfolded by auto with A3 have T2: "m ∈ \<int>" "n ∈ \<int>" "?p ∈ \<int>" "?q ∈ \<int>" "?r ∈ \<int>" using PositiveSet_def by auto from A2 A3 have T3: "?r\<rs>\<one> ∈ \<int>⇩_{+}" "?p\<rs>\<one> ∈ \<int>⇩_{+}" "?q\<rs>\<one> ∈ \<int>⇩_{+}" using pos_int_closed_add_unfolded by auto from A1 A3 have VII: "m\<ra>n \<lsq> f`(?r)" "m \<lsq> f`(?p)" "n \<lsq> f`(?q)" using Int_ZF_2_4_L2 pos_int_closed_add_unfolded by auto from A1 A3 T3 have VIII: "f`(?r\<rs>\<one>) \<lsq> m\<ra>n" "f`(?p\<rs>\<one>) \<lsq> m" "f`(?q\<rs>\<one>) \<lsq> n" using pos_int_closed_add_unfolded Int_ZF_2_4_L4 by auto have "f`(?r\<rs>?p\<rs>?q) \<lsq> L\<ra>M" proof - from IV T2 have "f`(?r\<rs>?p\<rs>?q) \<lsq> f`(?r)\<rs>f`(?p)\<rs>f`(?q)\<ra>M" by simp moreover from I II T2 VIII have "f`(?r) \<lsq> f`(?r\<rs>\<one>) \<ra> L" "f`(?r\<rs>\<one>) \<ra> L \<lsq> m\<ra>n\<ra>L" using int_ord_transl_inv by auto then have "f`(?r) \<lsq> m\<ra>n\<ra>L" by (rule Int_order_transitive) with VII have "f`(?r) \<rs> f`(?p) \<lsq> m\<ra>n\<ra>L\<rs>m" using int_ineq_add_sides by simp with I T2 VII have "f`(?r) \<rs> f`(?p) \<rs> f`(?q) \<lsq> n\<ra>L\<rs>n" using Int_ZF_1_2_L9 int_ineq_add_sides by simp with I III T2 have "f`(?r) \<rs> f`(?p) \<rs> f`(?q) \<ra> M \<lsq> L\<ra>M" using Int_ZF_1_2_L3 int_ord_transl_inv by simp ultimately show "f`(?r\<rs>?p\<rs>?q) \<lsq> L\<ra>M" by (rule Int_order_transitive) qed moreover have "(\<rm>L)\<rs>L \<ra>K \<lsq> f`(?r\<rs>?p\<rs>?q)" proof - from I II T2 VIII have "f`(?p) \<lsq> f`(?p\<rs>\<one>) \<ra> L" "f`(?p\<rs>\<one>) \<ra> L \<lsq> m \<ra>L" using int_ord_transl_inv by auto then have "f`(?p) \<lsq> m \<ra>L" by (rule Int_order_transitive) with VII have "m\<ra>n \<rs>(m\<ra>L) \<lsq> f`(?r) \<rs> f`(?p)" using int_ineq_add_sides by simp with I T2 have "n \<rs> L \<lsq> f`(?r) \<rs> f`(?p)" using Int_ZF_1_2_L9 by simp moreover from I II T2 VIII have "f`(?q) \<lsq> f`(?q\<rs>\<one>) \<ra> L" "f`(?q\<rs>\<one>) \<ra> L \<lsq> n \<ra>L" using int_ord_transl_inv by auto then have "f`(?q) \<lsq> n \<ra>L" by (rule Int_order_transitive) ultimately have "n \<rs> L \<rs> (n\<ra>L) \<lsq> f`(?r) \<rs> f`(?p) \<rs> f`(?q)" using int_ineq_add_sides by simp with I V T2 have "(\<rm>L)\<rs>L \<ra>K \<lsq> f`(?r) \<rs> f`(?p) \<rs> f`(?q) \<ra> K" using Int_ZF_1_2_L3 int_ord_transl_inv by simp moreover from VI T2 have "f`(?r) \<rs> f`(?p) \<rs> f`(?q) \<ra> K \<lsq> f`(?r\<rs>?p\<rs>?q)" by simp ultimately show "(\<rm>L)\<rs>L \<ra>K \<lsq> f`(?r\<rs>?p\<rs>?q)" by (rule Int_order_transitive) qed ultimately show "f`(?r\<rs>?p\<rs>?q) \<lsq> L\<ra>M ∧ (\<rm>L)\<rs>L\<ra>K \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" by simp qed } ultimately show "∃U∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n)) \<lsq> U" "∃N∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. N \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" by auto qed text{*The expression $f^{-1}(m+n)-f^{-1}(m) - f^{-1}(n)$ is uniformly bounded for all pairs $\langle m,n \rangle \in$ @{text "\<int>⇩_{+}×\<int>⇩_{+}"}. Recall that in the @{text "int1"} context @{text "ε(f,x)"} is defined so that $\varepsilon(f,\langle m,n \rangle ) = f^{-1}(m+n)-f^{-1}(m) - f^{-1}(n)$.*} lemma (in int1) Int_ZF_2_4_L8: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "∀m∈\<int>⇩_{+}. f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" shows "∃M. ∀x∈\<int>⇩_{+}×\<int>⇩_{+}. abs(ε(f,x)) \<lsq> M" proof - from A1 A2 have "∃U∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n)) \<lsq> U" "∃N∈\<int>. ∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. N \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" using Int_ZF_2_4_L7 by auto then obtain U N where I: "∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n)) \<lsq> U" "∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. N \<lsq> f`(f¯(m\<ra>n)\<rs>f¯(m)\<rs>f¯(n))" by auto have "\<int>⇩_{+}×\<int>⇩_{+}≠ 0" using int_one_two_are_pos by auto moreover from A1 have "f: \<int>->\<int>" using AlmostHoms_def by simp moreover from A1 have "∀a∈\<int>.∃b∈\<int>⇩_{+}.∀x. b\<lsq>x --> a \<lsq> f`(x)" using Int_ZF_2_3_L5 by simp moreover from A1 have "∀a∈\<int>.∃b∈\<int>⇩_{+}.∀y. b\<lsq>y --> f`(\<rm>y) \<lsq> a" using Int_ZF_2_3_L5A by simp moreover have "∀x∈\<int>⇩_{+}×\<int>⇩_{+}. ε(f,x) ∈ \<int> ∧ f`(ε(f,x)) \<lsq> U ∧ N \<lsq> f`(ε(f,x))" proof - { fix x assume A3: "x ∈ \<int>⇩_{+}×\<int>⇩_{+}" let ?m = "fst(x)" let ?n = "snd(x)" from A3 have T: "?m ∈ \<int>⇩_{+}" "?n ∈ \<int>⇩_{+}" "?m\<ra>?n ∈ \<int>⇩_{+}" using pos_int_closed_add_unfolded by auto with A1 have "f¯(?m\<ra>?n) ∈ \<int>" "f¯(?m) ∈ \<int>" "f¯(?n) ∈ \<int>" using Int_ZF_2_4_L2 PositiveSet_def by auto with I T have "ε(f,x) ∈ \<int> ∧ f`(ε(f,x)) \<lsq> U ∧ N \<lsq> f`(ε(f,x))" using Int_ZF_1_1_L5 by auto } thus ?thesis by simp qed ultimately show "∃M.∀x∈\<int>⇩_{+}×\<int>⇩_{+}. abs(ε(f,x)) \<lsq> M" by (rule Int_ZF_1_6_L4) qed text{*The (candidate for) inverse of a positive slope is a (well defined) function on @{text "\<int>⇩_{+}"}.*} lemma (in int1) Int_ZF_2_4_L9: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "g = {⟨p,f¯(p)⟩. p∈\<int>⇩_{+}}" shows "g : \<int>⇩_{+}->\<int>⇩_{+}" "g : \<int>⇩_{+}->\<int>" proof - from A1 have "∀p∈\<int>⇩_{+}. f¯(p) ∈ \<int>⇩_{+}" "∀p∈\<int>⇩_{+}. f¯(p) ∈ \<int>" using Int_ZF_2_4_L2 PositiveSet_def by auto with A2 show "g : \<int>⇩_{+}->\<int>⇩_{+}" and "g : \<int>⇩_{+}->\<int>" using ZF_fun_from_total by auto qed text{*What are the values of the (candidate for) the inverse of a positive slope?*} lemma (in int1) Int_ZF_2_4_L10: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "g = {⟨p,f¯(p)⟩. p∈\<int>⇩_{+}}" and A3: "p∈\<int>⇩_{+}" shows "g`(p) = f¯(p)" proof - from A1 A2 have "g : \<int>⇩_{+}->\<int>⇩_{+}" using Int_ZF_2_4_L9 by simp with A2 A3 show "g`(p) = f¯(p)" using ZF_fun_from_tot_val by simp qed text{*The (candidate for) the inverse of a positive slope is a slope.*} lemma (in int1) Int_ZF_2_4_L11: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "∀m∈\<int>⇩_{+}. f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" and A3: "g = {⟨p,f¯(p)⟩. p∈\<int>⇩_{+}}" shows "OddExtension(\<int>,IntegerAddition,IntegerOrder,g) ∈ \<S>" proof - from A1 A2 have "∃L. ∀x∈\<int>⇩_{+}×\<int>⇩_{+}. abs(ε(f,x)) \<lsq> L" using Int_ZF_2_4_L8 by simp then obtain L where I: "∀x∈\<int>⇩_{+}×\<int>⇩_{+}. abs(ε(f,x)) \<lsq> L" by auto from A1 A3 have "g : \<int>⇩_{+}->\<int>" using Int_ZF_2_4_L9 by simp moreover have "∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. abs(δ(g,m,n)) \<lsq> L" proof- { fix m n assume A4: "m∈\<int>⇩_{+}" "n∈\<int>⇩_{+}" then have "⟨m,n⟩ ∈ \<int>⇩_{+}×\<int>⇩_{+}" by simp with I have "abs(ε(f,⟨m,n⟩)) \<lsq> L" by simp moreover have "ε(f,⟨m,n⟩) = f¯(m\<ra>n) \<rs> f¯(m) \<rs> f¯(n)" by simp moreover from A1 A3 A4 have "f¯(m\<ra>n) = g`(m\<ra>n)" "f¯(m) = g`(m)" "f¯(n) = g`(n)" using pos_int_closed_add_unfolded Int_ZF_2_4_L10 by auto ultimately have "abs(δ(g,m,n)) \<lsq> L" by simp } thus "∀m∈\<int>⇩_{+}. ∀n∈\<int>⇩_{+}. abs(δ(g,m,n)) \<lsq> L" by simp qed ultimately show ?thesis by (rule Int_ZF_2_1_L24) qed text{*Every positive slope that is at least $2$ on positive integers almost has an inverse.*} lemma (in int1) Int_ZF_2_4_L12: assumes A1: "f ∈ \<S>⇩_{+}" and A2: "∀m∈\<int>⇩_{+}. f¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" shows "∃h∈\<S>. foh ∼ id(\<int>)" proof - let ?g = "{⟨p,f¯(p)⟩. p∈\<int>⇩_{+}}" let ?h = "OddExtension(\<int>,IntegerAddition,IntegerOrder,?g)" from A1 have "∃M∈\<int>. ∀n∈\<int>. f`(n) \<lsq> f`(n\<rs>\<one>) \<ra> M" using Int_ZF_2_1_L28 by simp then obtain M where I: "M∈\<int>" and II: "∀n∈\<int>. f`(n) \<lsq> f`(n\<rs>\<one>) \<ra> M" by auto from A1 A2 have T: "?h ∈ \<S>" using Int_ZF_2_4_L11 by simp moreover have "fo?h ∼ id(\<int>)" proof - from A1 T have "fo?h ∈ \<S>" using Int_ZF_2_1_L11 by simp moreover note I moreover { fix m assume A3: "m∈\<int>⇩_{+}" with A1 have "f¯(m) ∈ \<int>" using Int_ZF_2_4_L2 PositiveSet_def by simp with II have "f`(f¯(m)) \<lsq> f`(f¯(m)\<rs>\<one>) \<ra> M" by simp moreover from A1 A2 I A3 have "f`(f¯(m)\<rs>\<one>) \<ra> M \<lsq> m\<ra>M" using Int_ZF_2_4_L4 int_ord_transl_inv by simp ultimately have "f`(f¯(m)) \<lsq> m\<ra>M" by (rule Int_order_transitive) moreover from A1 A3 have "m \<lsq> f`(f¯(m))" using Int_ZF_2_4_L2 by simp moreover from A1 A2 T A3 have "f`(f¯(m)) = (fo?h)`(m)" using Int_ZF_2_4_L9 Int_ZF_1_5_L11 Int_ZF_2_4_L10 PositiveSet_def Int_ZF_2_1_L10 by simp ultimately have "m \<lsq> (fo?h)`(m) ∧ (fo?h)`(m) \<lsq> m\<ra>M" by simp } ultimately show "fo?h ∼ id(\<int>)" using Int_ZF_2_1_L32 by simp qed ultimately show "∃h∈\<S>. foh ∼ id(\<int>)" by auto qed text{* @{text "Int_ZF_2_4_L12"} is almost what we need, except that it has an assumption that the values of the slope that we get the inverse for are not smaller than $2$ on positive integers. The Arthan's proof of Theorem 11 has a mistake where he says "note that for all but finitely many $m,n\in N$ $p=g(m)$ and $q=g(n)$ are both positive". Of course there may be infinitely many pairs $\langle m,n \rangle$ such that $p,q$ are not both positive. This is however easy to workaround: we just modify the slope by adding a constant so that the slope is large enough on positive integers and then look for the inverse.*} theorem (in int1) pos_slope_has_inv: assumes A1: "f ∈ \<S>⇩_{+}" shows "∃g∈\<S>. f∼g ∧ (∃h∈\<S>. goh ∼ id(\<int>))" proof - from A1 have "f: \<int>->\<int>" "\<one>∈\<int>" "\<two> ∈ \<int>" using AlmostHoms_def int_zero_one_are_int int_two_three_are_int by auto moreover from A1 have "∀a∈\<int>.∃b∈\<int>⇩_{+}.∀x. b\<lsq>x --> a \<lsq> f`(x)" using Int_ZF_2_3_L5 by simp ultimately have "∃c∈\<int>. \<two> \<lsq> Minimum(IntegerOrder,{n∈\<int>⇩_{+}. \<one> \<lsq> f`(n)\<ra>c})" by (rule Int_ZF_1_6_L7) then obtain c where I: "c∈\<int>" and II: "\<two> \<lsq> Minimum(IntegerOrder,{n∈\<int>⇩_{+}. \<one> \<lsq> f`(n)\<ra>c})" by auto let ?g = "{⟨m,f`(m)\<ra>c⟩. m∈\<int>}" from A1 I have III: "?g∈\<S>" and IV: "f∼?g" using Int_ZF_2_1_L33 by auto from IV have "⟨f,?g⟩ ∈ AlEqRel" by simp with A1 have T: "?g ∈ \<S>⇩_{+}" by (rule Int_ZF_2_3_L9) moreover have "∀m∈\<int>⇩_{+}. ?g¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" proof fix m assume A2: "m∈\<int>⇩_{+}" from A1 I II have V: "\<two> \<lsq> ?g¯(\<one>)" using Int_ZF_2_1_L33 PositiveSet_def by simp moreover from A2 T have "?g¯(\<one>) \<lsq> ?g¯(m)" using Int_ZF_1_5_L3 int_one_two_are_pos Int_ZF_2_4_L5 by simp ultimately have "\<two> \<lsq> ?g¯(m)" by (rule Int_order_transitive) then have "\<two>\<rs>\<one> \<lsq> ?g¯(m)\<rs>\<one>" using int_zero_one_are_int Int_ZF_1_1_L4 int_ord_transl_inv by simp then show "?g¯(m)\<rs>\<one> ∈ \<int>⇩_{+}" using int_zero_one_are_int Int_ZF_1_2_L3 Int_ZF_1_5_L3 by simp qed ultimately have "∃h∈\<S>. ?goh ∼ id(\<int>)" by (rule Int_ZF_2_4_L12) with III IV show ?thesis by auto qed section{*Completeness*} text{*In this section we consider properties of slopes that are needed for the proof of completeness of real numbers constructred in @{text "Real_ZF_1.thy"}. In particular we consider properties of embedding of integers into the set of slopes by the mapping $m \mapsto m^S$ , where $m^S$ is defined by $m^S(n) = m\cdot n$.*} text{*If m is an integer, then $m^S$ is a slope whose value is $m\cdot n$ for every integer.*} lemma (in int1) Int_ZF_2_5_L1: assumes A1: "m ∈ \<int>" shows "∀n ∈ \<int>. (m⇧^{S})`(n) = m·n" "m⇧^{S}∈ \<S>" proof - from A1 have I: "m⇧^{S}:\<int>->\<int>" using Int_ZF_1_1_L5 ZF_fun_from_total by simp then show II: "∀n ∈ \<int>. (m⇧^{S})`(n) = m·n" using ZF_fun_from_tot_val by simp { fix n k assume A2: "n∈\<int>" "k∈\<int>" with A1 have T: "m·n ∈ \<int>" "m·k ∈ \<int>" using Int_ZF_1_1_L5 by auto from A1 A2 II T have "δ(m⇧^{S},n,k) = m·k \<rs> m·k" using Int_ZF_1_1_L5 Int_ZF_1_1_L1 Int_ZF_1_2_L3 by simp also from T have "… = \<zero>" using Int_ZF_1_1_L4 by simp finally have "δ(m⇧^{S},n,k) = \<zero>" by simp then have "abs(δ(m⇧^{S},n,k)) \<lsq> \<zero>" using Int_ZF_2_L18 int_zero_one_are_int int_ord_is_refl refl_def by simp } then have "∀n∈\<int>.∀k∈\<int>. abs(δ(m⇧^{S},n,k)) \<lsq> \<zero>" by simp with I show "m⇧^{S}∈ \<S>" by (rule Int_ZF_2_1_L5) qed text{*For any slope $f$ there is an integer $m$ such that there is some slope $g$ that is almost equal to $m^S$ and dominates $f$ in the sense that $f\leq g$ on positive integers (which implies that either $g$ is almost equal to $f$ or $g-f$ is a positive slope. This will be used in @{text "Real_ZF_1.thy"} to show that for any real number there is an integer that (whose real embedding) is greater or equal.*} lemma (in int1) Int_ZF_2_5_L2: assumes A1: "f ∈ \<S>" shows "∃m∈\<int>. ∃g∈\<S>. (m⇧^{S}∼g ∧ (f∼g ∨ g\<fp>(\<fm>f) ∈ \<S>⇩_{+}))" proof - from A1 have "∃m k. m∈\<int> ∧ k∈\<int> ∧ (∀p∈\<int>. abs(f`(p)) \<lsq> m·abs(p)\<ra>k)" using Arthan_Lem_8 by simp then obtain m k where I: "m∈\<int>" and II: "k∈\<int>" and III: "∀p∈\<int>. abs(f`(p)) \<lsq> m·abs(p)\<ra>k" by auto let ?g = "{⟨n,m⇧^{S}`(n) \<ra>k⟩. n∈\<int>}" from I have IV: "m⇧^{S}∈ \<S>" using Int_ZF_2_5_L1 by simp with II have V: "?g∈\<S>" and VI: "m⇧^{S}∼?g" using Int_ZF_2_1_L33 by auto { fix n assume A2: "n∈\<int>⇩_{+}" with A1 have "f`(n) ∈ \<int>" using Int_ZF_2_1_L2B PositiveSet_def by simp then have "f`(n) \<lsq> abs(f`(n))" using Int_ZF_2_L19C by simp moreover from III A2 have "abs(f`(n)) \<lsq> m·abs(n) \<ra> k" using PositiveSet_def by simp with A2 have "abs(f`(n)) \<lsq> m·n\<ra>k" using Int_ZF_1_5_L4A by simp ultimately have "f`(n) \<lsq> m·n\<ra>k" by (rule Int_order_transitive) moreover from II IV A2 have "?g`(n) = (m⇧^{S})`(n)\<ra>k" using Int_ZF_2_1_L33 PositiveSet_def by simp with I A2 have "?g`(n) = m·n\<ra>k" using Int_ZF_2_5_L1 PositiveSet_def by simp ultimately have "f`(n) \<lsq> ?g`(n)" by simp } then have "∀n∈\<int>⇩_{+}. f`(n) \<lsq> ?g`(n)" by simp with A1 V have "f∼?g ∨ ?g \<fp> (\<fm>f) ∈ \<S>⇩_{+}" using Int_ZF_2_3_L4C by simp with I V VI show ?thesis by auto qed text{*The negative of an integer embeds in slopes as a negative of the orgiginal embedding.*} lemma (in int1) Int_ZF_2_5_L3: assumes A1: "m ∈ \<int>" shows "(\<rm>m)⇧^{S}= \<fm>(m⇧^{S})" proof - from A1 have "(\<rm>m)⇧^{S}: \<int>->\<int>" and "(\<fm>(m⇧^{S})): \<int>->\<int>" using Int_ZF_1_1_L4 Int_ZF_2_5_L1 AlmostHoms_def Int_ZF_2_1_L12 by auto moreover have "∀n∈\<int>. ((\<rm>m)⇧^{S})`(n) = (\<fm>(m⇧^{S}))`(n)" proof fix n assume A2: "n∈\<int>" with A1 have "((\<rm>m)⇧^{S})`(n) = (\<rm>m)·n" "(\<fm>(m⇧^{S}))`(n) = \<rm>(m·n)" using Int_ZF_1_1_L4 Int_ZF_2_5_L1 Int_ZF_2_1_L12A by auto with A1 A2 show "((\<rm>m)⇧^{S})`(n) = (\<fm>(m⇧^{S}))`(n)" using Int_ZF_1_1_L5 by simp qed ultimately show "(\<rm>m)⇧^{S}= \<fm>(m⇧^{S})" using fun_extension_iff by simp qed text{*The sum of embeddings is the embeding of the sum.*} lemma (in int1) Int_ZF_2_5_L3A: assumes A1: "m∈\<int>" "k∈\<int>" shows "(m⇧^{S}) \<fp> (k⇧^{S}) = ((m\<ra>k)⇧^{S})" proof - from A1 have T1: "m\<ra>k ∈ \<int>" using Int_ZF_1_1_L5 by simp with A1 have T2: "(m⇧^{S}) ∈ \<S>" "(k⇧^{S}) ∈ \<S>" "(m\<ra>k)⇧^{S}∈ \<S>" "(m⇧^{S}) \<fp> (k⇧^{S}) ∈ \<S>" using Int_ZF_2_5_L1 Int_ZF_2_1_L12C by auto then have "(m⇧^{S}) \<fp> (k⇧^{S}) : \<int>->\<int>" "(m\<ra>k)⇧^{S}: \<int>->\<int>" using AlmostHoms_def by auto moreover have "∀n∈\<int>. ((m⇧^{S}) \<fp> (k⇧^{S}))`(n) = ((m\<ra>k)⇧^{S})`(n)" proof fix n assume A2: "n∈\<int>" with A1 T1 T2 have "((m⇧^{S}) \<fp> (k⇧^{S}))`(n) = (m\<ra>k)·n" using Int_ZF_2_1_L12B Int_ZF_2_5_L1 Int_ZF_1_1_L1 by simp also from T1 A2 have "… = ((m\<ra>k)⇧^{S})`(n)" using Int_ZF_2_5_L1 by simp finally show "((m⇧^{S}) \<fp> (k⇧^{S}))`(n) = ((m\<ra>k)⇧^{S})`(n)" by simp qed ultimately show "(m⇧^{S}) \<fp> (k⇧^{S}) = ((m\<ra>k)⇧^{S})" using fun_extension_iff by simp qed text{*The composition of embeddings is the embeding of the product.*} lemma (in int1) Int_ZF_2_5_L3B: assumes A1: "m∈\<int>" "k∈\<int>" shows "(m⇧^{S}) o (k⇧^{S}) = ((m·k)⇧^{S})" proof - from A1 have T1: "m·k ∈ \<int>" using Int_ZF_1_1_L5 by simp with A1 have T2: "(m⇧^{S}) ∈ \<S>" "(k⇧^{S}) ∈ \<S>" "(m·k)⇧^{S}∈ \<S>" "(m⇧^{S}) o (k⇧^{S}) ∈ \<S>" using Int_ZF_2_5_L1 Int_ZF_2_1_L11 by auto then have "(m⇧^{S}) o (k⇧^{S}) : \<int>->\<int>" "(m·k)⇧^{S}: \<int>->\<int>" using AlmostHoms_def by auto moreover have "∀n∈\<int>. ((m⇧^{S}) o (k⇧^{S}))`(n) = ((m·k)⇧^{S})`(n)" proof fix n assume A2: "n∈\<int>" with A1 T2 have "((m⇧^{S}) o (k⇧^{S}))`(n) = (m⇧^{S})`(k·n)" using Int_ZF_2_1_L10 Int_ZF_2_5_L1 by simp moreover from A1 A2 have "k·n ∈ \<int>" using Int_ZF_1_1_L5 by simp with A1 A2 have "(m⇧^{S})`(k·n) = m·k·n" using Int_ZF_2_5_L1 Int_ZF_1_1_L7 by simp ultimately have "((m⇧^{S}) o (k⇧^{S}))`(n) = m·k·n" by simp also from T1 A2 have "m·k·n = ((m·k)⇧^{S})`(n)" using Int_ZF_2_5_L1 by simp finally show "((m⇧^{S}) o (k⇧^{S}))`(n) = ((m·k)⇧^{S})`(n)" by simp qed ultimately show "(m⇧^{S}) o (k⇧^{S}) = ((m·k)⇧^{S})" using fun_extension_iff by simp qed text{*Embedding integers in slopes preserves order.*} lemma (in int1) Int_ZF_2_5_L4: assumes A1: "m\<lsq>n" shows "(m⇧^{S}) ∼ (n⇧^{S}) ∨ (n⇧^{S})\<fp>(\<fm>(m⇧^{S})) ∈ \<S>⇩_{+}" proof - from A1 have "m⇧^{S}∈ \<S>" and "n⇧^{S}∈ \<S>" using Int_ZF_2_L1A Int_ZF_2_5_L1 by auto moreover from A1 have "∀k∈\<int>⇩_{+}. (m⇧^{S})`(k) \<lsq> (n⇧^{S})`(k)" using Int_ZF_1_3_L13B Int_ZF_2_L1A PositiveSet_def Int_ZF_2_5_L1 by simp ultimately show ?thesis using Int_ZF_2_3_L4C by simp qed text{*We aim at showing that $m\mapsto m^S$ is an injection modulo the relation of almost equality. To do that we first show that if $m^S$ has finite range, then $m=0$.*} lemma (in int1) Int_ZF_2_5_L5: assumes "m∈\<int>" and "m⇧^{S}∈ FinRangeFunctions(\<int>,\<int>)" shows "m=\<zero>" using assms FinRangeFunctions_def Int_ZF_2_5_L1 AlmostHoms_def func_imagedef Int_ZF_1_6_L8 by simp text{*Embeddings of two integers are almost equal only if the integers are equal.*} lemma (in int1) Int_ZF_2_5_L6: assumes A1: "m∈\<int>" "k∈\<int>" and A2: "(m⇧^{S}) ∼ (k⇧^{S})" shows "m=k" proof - from A1 have T: "m\<rs>k ∈ \<int>" using Int_ZF_1_1_L5 by simp from A1 have "(\<fm>(k⇧^{S})) = ((\<rm>k)⇧^{S})" using Int_ZF_2_5_L3 by simp then have "m⇧^{S}\<fp> (\<fm>(k⇧^{S})) = (m⇧^{S}) \<fp> ((\<rm>k)⇧^{S})" by simp with A1 have "m⇧^{S}\<fp> (\<fm>(k⇧^{S})) = ((m\<rs>k)⇧^{S})" using Int_ZF_1_1_L4 Int_ZF_2_5_L3A by simp moreover from A1 A2 have "m⇧^{S}\<fp> (\<fm>(k⇧^{S})) ∈ FinRangeFunctions(\<int>,\<int>)" using Int_ZF_2_5_L1 Int_ZF_2_1_L9D by simp ultimately have "(m\<rs>k)⇧^{S}∈ FinRangeFunctions(\<int>,\<int>)" by simp with T have "m\<rs>k = \<zero>" using Int_ZF_2_5_L5 by simp with A1 show "m=k" by (rule Int_ZF_1_L15) qed text{*Embedding of $1$ is the identity slope and embedding of zero is a finite range function.*} lemma (in int1) Int_ZF_2_5_L7: shows "\<one>⇧^{S}= id(\<int>)" "\<zero>⇧^{S}∈ FinRangeFunctions(\<int>,\<int>)" proof - have "id(\<int>) = {⟨x,x⟩. x∈\<int>}" using id_def by blast then show "\<one>⇧^{S}= id(\<int>)" using Int_ZF_1_1_L4 by simp have "{\<zero>⇧^{S}`(n). n∈\<int>} = {\<zero>·n. n∈\<int>}" using int_zero_one_are_int Int_ZF_2_5_L1 by simp also have "… = {\<zero>}" using Int_ZF_1_1_L4 int_not_empty by simp finally have "{\<zero>⇧^{S}`(n). n∈\<int>} = {\<zero>}" by simp then have "{\<zero>⇧^{S}`(n). n∈\<int>} ∈ Fin(\<int>)" using int_zero_one_are_int Finite1_L16 by simp moreover have "\<zero>⇧^{S}: \<int>->\<int>" using int_zero_one_are_int Int_ZF_2_5_L1 AlmostHoms_def by simp ultimately show "\<zero>⇧^{S}∈ FinRangeFunctions(\<int>,\<int>)" using Finite1_L19 by simp qed text{*A somewhat technical condition for a embedding of an integer to be "less or equal" (in the sense apriopriate for slopes) than the composition of a slope and another integer (embedding).*} lemma (in int1) Int_ZF_2_5_L8: assumes A1: "f ∈ \<S>" and A2: "N ∈ \<int>" "M ∈ \<int>" and A3: "∀n∈\<int>⇩_{+}. M·n \<lsq> f`(N·n)" shows "M⇧^{S}∼ fo(N⇧^{S}) ∨ (fo(N⇧^{S})) \<fp> (\<fm>(M⇧^{S})) ∈ \<S>⇩_{+}" proof - from A1 A2 have "M⇧^{S}∈ \<S>" "fo(N⇧^{S}) ∈ \<S>" using Int_ZF_2_5_L1 Int_ZF_2_1_L11 by auto moreover from A1 A2 A3 have "∀n∈\<int>⇩_{+}. (M⇧^{S})`(n) \<lsq> (fo(N⇧^{S}))`(n)" using Int_ZF_2_5_L1 PositiveSet_def Int_ZF_2_1_L10 by simp ultimately show ?thesis using Int_ZF_2_3_L4C by simp qed text{*Another technical condition for the composition of a slope and an integer (embedding) to be "less or equal" (in the sense apriopriate for slopes) than embedding of another integer.*} lemma (in int1) Int_ZF_2_5_L9: assumes A1: "f ∈ \<S>" and A2: "N ∈ \<int>" "M ∈ \<int>" and A3: "∀n∈\<int>⇩_{+}. f`(N·n) \<lsq> M·n " shows "fo(N⇧^{S}) ∼ (M⇧^{S}) ∨ (M⇧^{S}) \<fp> (\<fm>(fo(N⇧^{S}))) ∈ \<S>⇩_{+}" proof - from A1 A2 have "fo(N⇧^{S}) ∈ \<S>" "M⇧^{S}∈ \<S>" using Int_ZF_2_5_L1 Int_ZF_2_1_L11 by auto moreover from A1 A2 A3 have "∀n∈\<int>⇩_{+}. (fo(N⇧^{S}))`(n) \<lsq> (M⇧^{S})`(n) " using Int_ZF_2_5_L1 PositiveSet_def Int_ZF_2_1_L10 by simp ultimately show ?thesis using Int_ZF_2_3_L4C by simp qed end